Welcome to the ICML 2024 Session Finder! This tool helps you search for sessions and ask questions about them using a Retrieval Augmented Generation (RAG) pattern. The selected sessions include the session name, authors, time, location, and abstract, which are added to the language model's prompt.
Access the ICML Session Finder here: Link
- Session Search: Search for ICML sessions using natural language queries.
- Selected Sessions: Add sessions to your list of selected sessions.
- Session Chat: Ask questions about your selected sessions and receive AI-generated responses.
- Set a Good Window Height: Adjust the chat window height in the sidebar.
- Ask Questions: Ask any questions about all the sessions to find relevant content.
- Utilize the Language Model: Request the language model to perform tasks on the selected sessions, such as "Make me a schedule for these" or "Tell me what these have in common."
- Enjoy ICML!
I participated in the Parlance Labs Conference - Mastering LLMs: A Conference For Developers & Data Scientists which gave me 1000s of dollars of free credits for all kinds of services including Modal Labs and OpenAI and I wanted to put them to good use and make something practical.
- Hosting: Modal Labs
- CI/CD: Github Actions
- LLM: OpenAI GPT-4o-mini
- Embeddings: OpenAI Text-003-large
- VectorDB: LanceDB
- Reranker: Cohere.ai
- Frontend: Streamlit
- Dataset Hosting: Huggingface Datasets
The dataset is downloaded from ICML Downloads
If you have any questions leave an issue on Github!
If you like this give me shoutout on Twitter or LinkedIn!
Last updated Thursday July 21st!
Double check the programme for correctness!
No warranty provided or any claims to the content.